Model profile
Llama 3 1 405B Instruct Turbo
1eval lineages
Evidence summary
Published evidence spans 1 evals and 3 of 7 behavior components. Its strongest relative result is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #94 of 260); its weakest is Enkrypt AI Safety Leaderboard (bias_attack_non_success_rate, #181 of 260).
Compare this model
Only models sharing at least one published sub-eval are listed.
Published eval results
Rank is within that sub-eval. Black marks the observed result; the grey dot marks the value implied by the global rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Better | Source |
|---|---|---|---|---|---|
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #181 / 260 | 12.92 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #94 / 260 | 90.17 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #121 / 260 | 73.33 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #136 / 258 | 96.05 | ↑ higher | Source ↗official |